classification of domain-specific bpmn extensions...performance measurement a own ext uml explicit...
TRANSCRIPT
Classification of Domain-Specific BPMN
Extensions
Richard Braun and Werner Esswein
Technische Universitat DresdenChair of Wirtschaftsinformatik, esp. Systems Development
01062 Dresden, Germany{richard.braun,werner.esswein}@tu-dresden.de
Abstract. BPMN is a standard for modeling business processes andprovides meta model concepts for the design of extensions. Thus, domain-specific extensions of the BPMN are facilitated. This research articleprovides an overview of BPMN extension development by the descrip-tive analysis and classification of 30 BPMN extensions. An extensiveliterature review was conducted in order to find published extensions.Further, a classification framework was designed to enable a comprehen-sive analysis of each extension. The analysis showed, that four out of fiveextensions are not compliant with the BPMN standard. Also, we foundseveral methodological shortcomings that should be tackled in furtherresearch.
Keywords: BPMN Extensions, Domain-specific Extension, ModelingLanguages, DSML, DSML Repository, Business Process Modeling.
1 Introduction and Motivation
The Business Process Model and Notation (BPMN) is an ISO standard for mod-eling business processes and a de-facto standard in professional practice [1], [2].BPMN provides a set of generic business process elements, independent from aspecific domain. However, it is often necessary to extend BPMN with individ-ual concepts in order to represent characteristics of a particular domain (e.g.,health care or security management). On the one hand, such domain-specificaspects can be integrated within a dedicated domain-specific modeling language(DSML) [3], [4]. On the other hand, BPMN can be extended with domain-specificconcepts in order to reuse the modeling language, take advantage of its bene-fits (e.g., standardization, tool support) and avoid expensive development of aDSML from the scratch. This research article investigates the current state ofthe art of BPMN extension development. A BPMN extension is understood asthe enhancement of functionality of the BPMN, following the extension mecha-nism defined in specification. In its own, the standard-conform BPMN extensionis neither useful nor functional (referring to [5], [6]).
U. Frank et al. (Eds.): PoEM 2014, LNBIP 197, pp. 42–57, 2014.c© IFIP International Federation for Information Processing 2014
Classification of Domain-Specific BPMN Extensions 43
1.1 BPMN Extensibility
BPMN is one of very few modeling languages that provides generic extensionelements within the meta model that enables the definition of domain-specificlanguage extensions [7]. BPMN provides an extension by addition mechanismthat ensures the validity of the BPMN core elements ([8], [7], p. 44). The fol-lowing elements are defined for the specification of valid BPMN extensions: AnExtension Definition is a named group of new attributes which can be usedby BPMN elements. Thus, new elements can be built implicitly. An ExtensionDefinition consists of several Extension Attribute Definitions that define theparticular attributes. Values of these Extension Attribute Definitions can be de-fined by the Extension Attribute Value class. Therefore, primitive types from theMeta Object Facility can be used [7]. The element Extension binds the entireextension definition and its attributes to a BPMN model definition. By doingso, all extension elements are accessible for existing BPMN elements ([7], p. 58).Further, external relationships can used for the integration of BPMN artifactsand UML elements, for instance (see [7], p. 62). Despite the fact that BPMN pro-vides a well-defined extension interface, a process model for the straightforwarddevelopment of extensions is missing. To the best of our knowledge, there is onlyone research article addressing this problem: [8] defines a model-transformationbased procedure model for the methodical development of valid BPMN exten-sions models based on conceptual domain models. However, the approach lacksin terms of a detailed analysis and consideration of the domain since it is a moreengineering driven approach that aims to provide clear transformation rules.Therefore, [9] extends the method with regard to the domain analysis and out-line several preceded steps in order to conceptualize the domain and identify areasoned need for extension.
1.2 Research Objective
As stated above, a detailed process model for the application of these extensionelements is missing and the development of an extension remains more or less“ad hoc”. Especially from a design science perspective, this lack of rigor is insuf-ficient (e.g., [10]). For example, there is neither guidance in terms of the domainconceptualization nor a semantic analysis between a specific domain conceptand BPMN elements. The mentioned approaches address this issue, but eitherlack in terms of domain analysis [8] or level of detail and applicability [9]. Weargue, that it is crucial to evolve a holistic process model for BPMN extensiondevelopment to ensure standard conformity, comprehensibility and falsifiability.Therefore, it is unavoidable to gain a comprehensive overview of the state ofthe art in the context of BPMN extensions. Thus, this research article aims toprovide a systematic, descriptive analysis of BPMN extensions in order to giveindications of both methodological and domain-specific aspects within BPMNextension development.
44 R. Braun and W. Esswein
1.3 Research Method
In order to find published BPMN extensions, a systematic literature reviewwas conducted. We have applied the method of [12] that was configured asfollows: The scope of our review was a broad analysis of BPMN extensions.According to [11] the review is conceptual, has a research outcome focus, aims tointegrate existing results (to a classification schema), has an exhaustive coverageand addresses a general audience. Second, the topic was conceptualized by thedefinition of relevant search phrases and keywords (see [12]), such as “BPMNextension”, “extend BPMN”, “enhance BPMN”, “extending BPMN”, “domain-specifc BPMN” and “domain BPMN”. Third, the literature search process wasconducted [12]. Therefore, the journal and conference list of the german researchorganizationWKWI was used [13]. Also, literature databases and search engineslike Google Scholar, Springer Link, Science Direct, AIS Digital Library and theIEEE Xplore Digital Library were used. Besides, each found article was usedfor a backward search. This search procedure resulted in a set of 39 articles,whose content were reviewed. Publications, focusing on early BPMN extensionsthat are now part of the language (e.g., [14]) or articles that did not provideany conceptual advices on their extension (e.g., [15], [16]), were discarded and aset of 30 articles remain for in-detail analysis that was conducted subsequently.Therefore, a multi-perspective analysis framework has been designed in orderto facilitate a comparison of the identified extensions. The systematization ofall BPMN extensions and the derivation of the state of the art represent thesynthesis of the review process. Finally, research gaps and aspects for furtherresearch were derived.
The structure of the article is as follows. Section 2 presents the extensionanalysis framework containing four main classes and all relevant criteria. Insection 3, the results of the literature review are analyzed within the framework.Section 4 provides indications as a result of the classification. The article endswith a short summary.
2 BPMN Extension Analysis Framework
The reasonable analysis of BPMN extensions requires the definition of a descrip-tion framework. In the context of BPMN, there are no comparable approaches,that could be leveraged for the derivation of such criteria. However, there arefew research articles addressing a systematic overview or classification of exten-sions in the field of the workflow modeling language BPEL [6] and UML profiles[17]. [6] evolves a classification framework for BPEL extensions based on theanalysis of 62 publications. Since their work focusses on workflow aspects, thereuse of the entire classification framework is not reasonable. Nevertheless, somecriteria like standard conformity, extension purpose and basic characteristics areadapted in the context of BPMN. [17] provides a systematic review of UMLprofiles based on the analysis of 39 publications. Although the focus lies on theanalysis of UML profiles, the consideration of extended meta classes (see [17],p. 413) is promising in the context of BPMN since both modeling languages are
Classification of Domain-Specific BPMN Extensions 45
defined by the Meta Object Facility (MOF). Referring to the mentioned worksand the research objective of this paper, the following classes for descriptiveanalysis of BPMN extensions were defined: “Basic attributes”, “standard con-formity”, “applied method” and “extension”. Each class, its containing criteriaand all classification values are described in the tables subsequently. If necessary,detailed explanations of single criteria are given.
Table 1. Basic Attributes
Criterion Description ValuesAuthors Authors of the publication (reference)Year Year of publication 2007 - 2014Version Affected BPMN version BPMN 1.x; BPMN 2.0 (since 2011)Medium Publication medium J (journal); P (proceedings); O (others)Title Title of the extension e.g., BPMN4WSNDomain Affected domain or area of discourse e.g., Artifacts or ResourcesPurpose Derived purpose D (descriptive); A (analytic); E (execu-
tion)
The criterion Domain describes the affected domain, the application fields orthe general area of discourse of the extension. During analysis, similar domains(e.g., Security Management and Risk Management) were merged to single do-mains (e.g., Risk Management) in order to consolidate them. Criterion Purposestands for the primarily purpose of the extension. An extension was classified as“descriptive” (D) if its focus lies on the description of a domain. It was classifiedas “analytic” (A) if the main purpose consists in facilitating some kind of analy-sis of existing BPMN models. If the extension aims to support process execution(e.g., supporting domain-specific transformation to BPEL), the extension wasclassified as “execution” (E).
Table 2. Standard Conformity
Criterion Description ValuesDefinition Type of extension definition Valid Ext; Own Ext; Own Ext Nota-
tion; NoneAbstract Syntax Definition of the meta model e.g., UML, Ext MM (BPMN exten-
sion meta model)Concrete Syntax Definition of new notations explicit; implicit (by example); noneSemantic Conflicts Are there any semantic conflicts with
the BPMN standard?no; yes
The “Standard Conformity” class contains criteria regarding the syntacticaland semantic correctness of the extension in the light of the BPMN standard(see section 1). Criterion Definition describes the way the extension is definedand explicated. “Valid Ext” stands for the definition as BPMN extension model.“Own Ext” outlines the application of a dedicated definition (e.g., UML model).“Own Ext Notation” stands for a solely graphical definition (e.g., by new icons).
46 R. Braun and W. Esswein
Table 3. Method
Criterion Description ValuesRequirements Anal-ysis
Is there any analysis or consideration of require-ments to the extension?
explicit; implicit; no
Semantic Fit Check Is there any discussion of the semantic fit of do-main concepts with BPMN elements for the iden-tification of extension need?
yes; partly; no
Reuse of Artifacts Many domains already provide some artifactssuch as ontologies. The reuse and integration ofthem might be useful.
yes; partly; no
Process Model Is any methodological approach applied (if yes,which one)?
Stroppi et al.; BPMNext; yes (own); no
Further, the definition of customized or new graphical elements is considered bythe criterion Concrete Syntax. Also, we have analyzed whether a single extensioncontains obvious semantic conflicts.
As stated at the beginning of the paper, the methodological development ofBPMN extensions is important, but BPMN standard does not provide any guid-ance and only very few publications addressing this topic. Thus, both method-ological and domain-analysis aspects are investigated within the class “Method”.For instance, requirements analysis is perceived as essential for the developmentof artifacts. It might be reasonable to reuse existing domain artifacts for reasonsof redundancy and communication with domain experts. Also, a discussion ofthe semantic fit with BPMN elements is necessary to constitute the need forextension elements.
The class “Extension” describes all extensions and customizations for the in-tegration of domain-specific aspects in BPMN. The first part contains all newlyadded elements, relations, properties and diagrams. Therefore, first of all it wasanalyzed whether the extension was defined by a meta model. If not, we have
Table 4. Extension
Criterion Description ValuesNew elements
Elements New elements and enumerations (up to threeexample elements are stated)
(individual)
Count Number of new elements (if the number isin brackets, a meta model is missing and theelements are derived logically; e.g., [18], [19])
(individual)
Size Class Derived extension size class, based on thenumber of extension elements
Heavy (>17); large (11-17);light (6-10); tiny (<6)
Diagrams Does the extension provide a new diagram? yes; noExtended or customized elements
Relations Extending a BPMN element by new naviga-ble relations to or from the element
BPMN element(s)
Properties New owned properties of a BPMN element BPMN element(s)Specialization Adding new sub classes to a BPMN element BPMN element(s)Enhancement Adding a new super class to a BPMN element BPMN element(s)Graphical Custom. Specifying a BPMN element by a new graph-
ical representation (see [7], p. 44)BPMN element(s)
Count Number of extended elements (individual)Extension Style Identified extension styles Codes from table 5
Classification of Domain-Specific BPMN Extensions 47
Table 5. Extension Styles
Code Name DescriptionAbstract Syntax
AS-Sp Specialization Specialization of elements by inheriting from the standard el-ement and extending it (e.g., by additional properties).
AS-A Additive (various) Set of both new elements and new relations or properties(both optional and mandatory). Thus, the meta model ex-tension is largely integrated within the BPMN meta model.
AS-A-B Additive (block) Set of new elements that is related to the BPMN core model byonly one or two relationships. Thus, the meta model extensionlooks like a well definable extension block.
AS-En Enumeration Domain-specific ranges in the form of enumeration elements.Semantics
Sem-Co Concretisation Specification of under specified elements (e.g., Lanes [7]).Sem-Ch Change Dedicated change of some element’s semantics, which is not
permitted within BPMN.Concrete Syntax
CS-Dg Diagram, view Adding a new diagram or view to BPMN (e.g., resource dia-gram as complement of the collaboration diagram).
CS-Cu Customization Customization of graphical elements (e.g., data objects).CS-Co Color Color highlighting of elements or parts with special semantics.CS-Ah Ad hoc Elusive definition of an extension by graphical icons, without
any abstract syntax.
tried to identify new elements based on explanations in the research article. Eventhough these explanations were missing, we looked for new, solely graphicallydefined elements (see Graphical Custom.). Criterion Size Class is a simple pa-rameter for the number of new elements 1. Further, the so-called extension stylesof an extension were analyzed in order to get a better understanding of the wayan intended extension was implemented and expressed. Therefore, ten extensionstyles were derived from the set of all 30 extensions inductively. Each extensionstyle is assigned to one of the following classes that were adapted from methodengineering: Abstract syntax, semantics and concrete syntax. Table 5 presentsand describes all styles in detail. Each analyzed extension can have multipleextension styles.
3 BPMN Extension Classification
The conducted literature review resulted in a set of 30 BPMN extensions. Eachextension was analyzed with respect to the abovementioned framework. Figure1 presents the results of the analysis regarding to basic attributes, conformity tothe standard and the applied method. Figure 6 presents the results regarding tothe syntactical definition of the extensions.
3.1 Basic Extension Attributes
Themajority of the considered extensions is related to BPMNversion 2.0 (76,6%).Extensions are mainly published in conference proceedings (60,0%) or as research
1 Size classes were generated by the application of the k-means algorithm over allelement counts (k=4; euclidean distance).
48 R. Braun and W. Esswein
Bas
ic A
ttrib
utes
St
anda
rd C
onfo
rmity
M
etho
d
Authors
Year
Version
Medium
Title
Domain
Purpose
Definition
Abstract Syntax
Concrete Syntax
Semantic Conflicts
Req. Analysis
Semantic Fit Check
Reuse of Artifacts
Process Model
Altu
hhov
et a
l. 20
13 2
.0
J N
o tit
le (S
ecur
ity R
isk
Man
agem
ent)
Ris
k M
anag
emen
t D
O
wn
Ext
N
otat
ion
UM
L E
xplic
it N
o im
plic
it (IS
SR
M)
yes
yes
(ISS
RM
) no
Aw
ad e
t al.
20
09 2
.0 O
N
o tit
le (R
esso
urce
Allo
catio
n C
onst
rain
ts)
Res
sour
ce
E
Ow
n E
xt
UM
L,
OC
L N
o N
o im
plic
it (p
atte
rns)
no
no
no
Boc
ciar
elli
& D
'Am
brog
io
2011
1.x
P
P
yBP
MN
P
erfo
rman
ce M
easu
rem
ent
A
Ow
n E
xt
UM
L,
OC
L N
o N
o no
no
ye
s
(MA
RTE
) (y
es)
Bra
mbi
lla e
t al.
2012
2.0
P
N
o tit
le (S
ocia
l BP
M)
Soc
ial B
PM
D
O
wn
Ext
(U
ML)
B
y ex
ampl
e N
o no
no
no
no
B
raun
& E
ssw
ein
2014
2.0
P
N
o tit
le (R
esso
urce
s in
E
ngin
eerin
g)
Res
ourc
e D
V
alid
Ext
E
xt M
M
Exp
licit
No
impl
icit
(ont
olog
y)
yes
yes
(Res
ML)
ye
s: e
xt. o
f Stro
ppi
et a
l. B
raun
et a
l.
2014
2.0
O
BP
MN
4CP
E
-Hea
lth
D
Val
id E
xt
Ext
MM
B
y ex
ampl
e N
o ex
plic
it ye
s ye
s (r
equi
rem
ents
) ye
s: e
xt. o
f Stro
ppi
et a
l. B
ruck
er e
t al.
2012
2.0
P
S
ecur
eBP
MN
R
isk
Man
agem
ent
D
Non
e N
o B
y ex
ampl
e N
o no
no
ye
s
(Sec
ureB
PM
) no
Cha
rfi e
t al.
20
10 2
.0
J A
O4B
PM
N
Asp
ect M
odel
ling
E
Non
e N
o B
y ex
ampl
e N
o no
no
pa
rtly
(A
OP
ele
men
ts)
no
Frie
dens
tab
et a
l.
2012
2.0
P
N
o tit
le (B
usin
ess
Act
ivity
M
onito
ring)
P
erfo
rman
ce M
easu
rem
ent
A
Ow
n E
xt
UM
L E
xplic
it N
o im
plic
it (d
escr
iptiv
e)
no
no
no
Gag
ne &
Tru
del
2009
1.x
P
Ti
me-
BP
MN
Ti
me
A
Non
e N
o B
y ex
ampl
e N
o im
plic
it (d
escr
iptiv
e)
no
yes
(tim
e as
pect
s)
no
Gro
ßkop
f 20
07 2
.0 O
N
o tit
le (R
esso
urce
In
form
atio
n La
yer)
R
esou
rces
E
N
one
UM
L N
o N
o no
ye
s no
no
Kop
p et
al.
2012
2.0
J
BP
MN
4TO
SC
A
Clo
ud
D
Non
e N
o E
xplic
it N
o ex
plic
it no
ye
s no
K
orhe
rr &
Lis
t 20
07 1
.x
J N
o tit
le (P
erfo
rman
ce
Mea
sure
s)
Per
form
ance
Mea
sure
men
t A
O
wn
Ext
U
ML
By
exam
ple
No
impl
icit
(des
crip
tive)
no
no
no
Lodh
i et a
l. 20
11 2
.0 O
N
o tit
le (P
roce
ss E
valu
atio
n)
Per
form
ance
Mea
sure
men
t A
O
wn
Ext
N
otat
ion
(UM
L)
By
exam
ple
(No)
no
no
pa
rtly
(eva
l. co
ncep
ts)
no
Lohm
ann
& N
yolt
2011
2.0
P
N
o tit
le (A
rtifa
ct-c
entri
c M
odel
ing)
A
rtifa
cts
D
Non
e N
o B
y ex
ampl
e N
o no
no
no
no
Mag
nani
& M
onte
si
2007
1.x
P
N
o tit
le (C
osts
) P
erfo
rman
ce M
easu
rem
ent
A
Non
e N
o B
y ex
ampl
e N
o no
no
no
no
M
agna
ni &
Mon
tesi
20
09 1
.x
O
BP
DM
N
Arti
fact
s A
O
wn
Ext
N
otat
ion
No
Exp
licit
Yes
im
plic
it (fe
atur
e co
mpa
rison
) no
ye
s (d
ata
flow
, ER
M)
no
Mar
cink
owsk
i & K
ucia
pski
201
2 2.
0 J
No
title
(Ris
k H
andl
ing)
R
isk
Man
agem
ent
D
Ow
n E
xt
UM
L B
y ex
ampl
e N
o no
no
no
no
M
ülle
r-W
icko
p &
Sch
ultz
20
13 2
.0
P
No
title
(Pro
cess
Aud
its)
Com
plia
nce
and
Aud
its
D
Val
id E
xt
Ext
MM
E
xplic
it N
o ex
plic
it no
no
(y
es)
BP
MN
ext
. N
atsc
hläg
er
2011
2.0
P
D
eont
ic B
PM
N
Deo
ntic
Ana
lysi
s A
N
one
No
No
No
no
no
yes
(deo
ntic
item
s)
no
Pill
at e
t al.
2012
2.0
P
B
PM
Nt
Sof
twar
e D
evel
opm
ent
D
Ow
n E
xt
Ext
MM
N
o N
o im
plic
it (d
escr
iptio
n)
no
yes
(S
PE
M 2
.0)
(yes
)
Rod
rigue
z et
al.
2007
1.x
J
No
title
(Sec
urity
) R
isk
Man
agem
ent
D
Ow
n E
xt
UM
L,
OC
L E
xplic
it N
o no
pa
rtly
yes
(req
uire
men
ts)
no
Sae
edi e
t al.
2010
2.0
P
N
o tit
le (S
eriv
ce Q
ualit
y R
equi
rem
ents
) Q
ualit
y M
anag
emen
t D
V
alid
Ext
E
xt M
M
By
exam
ple
No
impl
icit
(des
crip
tion)
no
no
(y
es)
BP
MN
ext
. S
alee
m &
Has
san
2012
2.0
J
No
title
(Sec
urity
R
equi
rem
ents
in S
OA
) R
isk
Man
agem
ent
D
Ow
n E
xt
UM
L E
xplic
it N
o no
no
no
no
Sch
leic
her e
t al.
2010
2.0
P
N
o tit
le (C
ompl
ianc
e R
equi
rem
ents
) C
ompl
ianc
e an
d A
udits
D
N
one
No
No
No
no
no
no
no
Stro
ppi e
t al.
2011
2.0
P
N
o tit
le (R
esso
urce
s)
Res
ourc
es
E
Val
id E
xt
Ext
MM
B
y ex
ampl
e N
o no
no
no
(y
es)
BP
MN
ext
. S
ungu
r et a
l. 20
13 2
.0
P
BP
MN
4WS
N
Sen
sors
D
O
wn
Ext
U
ML
Exp
licit
No
expl
icit
no
no
no
Sup
ulni
ece
et a
l. 20
12 2
.0
P
No
title
(kno
wle
dge)
K
now
ledg
e D
N
one
No
Exp
licit
Yes
ex
plic
it no
ye
s
(KM
DL)
no
Wol
ter &
Sch
aad
2007
1.x
P
N
o tit
le (A
utho
rizat
ion
Con
stra
ins)
A
utho
rizat
ion
A
Ow
n E
xt
UM
L B
y ex
ampl
e N
o no
no
no
no
Zor e
t al.
2011
2.0
P
N
o tit
le (M
anuf
actu
ring)
M
anuf
actu
ring
D
Non
e N
o E
xplic
it Y
es
no
partl
y no
no
Fig. 1. Analysis of the extensions regarding their basic attributes, BPMN standardconformity and the applied extension approach or method
Classification of Domain-Specific BPMN Extensions 49
J, 23,3%
P, 60,0%
O, 16,7%
D, 56,7%
A, 30,0%
E, 13,3%
0 1 2 3 4 5 6
Medium Purpose Domain
Fig. 2. Distribution of the observed extension attributes regarding the publicationmedium, the particular purpose and the addressd domain
reports (16,7%). Only every fourth extension is published in a journal, which couldbe interpreted as a lack of maturity in BPMN extension research (see figure 2).Also, we could not find any advice for cumulated research on single extensions.More than half of the publications reveal a descriptive purpose (56,7%) that aimsto describe some domain (e.g., sensor networks [20]). 30% of the extensions aimto enrich BPMN for specific analytical purposes such as process cost [21]. 13,3%focuses run-time or execution-oriented issues like resource allocation constraints[22]. The targeted domains of the extensions are very heterogenous; altogether17 domains were identified. Five publications address performance measurement[23], [24], [21], [25], [26], another five publications deal with issues related to riskmanagement [27], [28], [29], [30], [31] and four extensions are designed for resourcerelated issues [22], [32], [33], [34].
3.2 Standard Conformity
It is remarkable, that only 16,7% of the extensions are defined by the BPMNextension mechanism (see the first piechart in figure 3). Thus, four out of fiveextensions are not compliant with the BPMN meta model! These extensionsare either defined by a dedicated meta modeling approach (36,7%) using UMLor OCL expressions (e.g., [24], [23], [30]). Or these extensions do not have anymeta model and are defined solely by new notation elements (10,0%) like [35].36,7% of the extensions do not present any definition! It has to be stated, thatBPMN extension mechanism was introduced in version 2.0 in January 2011.Thus, actually all eleven extensions published before 2011 could not have anymethodical support. However, the consideration of the 19 extensions publishedafter 2011 reveals that only 21% were defined as BPMN extension meta modelsand still 32% do not provide any structured definition. It became obvious thatthe majority of extensions is not compliant with the BPMN standard.
Modeling language extensions generally requires the definition of customizedor added notation elements (see [7], p. 44). 40% of the analyzed extensionspresent the extended concrete syntax be describing new graphics explicitly. Other40% of the articles present new graphical elements implicitly within demon-stration models. 20% of the extensions do not define or explicate any kind of
50 R. Braun and W. Esswein
graphical extension. Further, BPMN specification claims to not contradict thesemantics of any BPMN element. Within the analysis process, not every part ofeach meta model was checked due to resource limitations and due to the fact thatmost of the articles were peer-reviewed before publication. However, we foundsemantic discrepancies in four extensions: [25] uses Pools and Lanes in order toexpress performances, although these elements are designated for organizationalunits, responsibilities or roles. [35] integrates data objects within the sequenceflow, although they must not have any direct effect on it. In a similar way, [36]integrates non-flow elements within the sequence flow what is not permitted.[37] specializes gateways to material gateways and use them for material trans-formations what is not the scope of gateways. Despite these few irregularities,the majority of the extensions do not contain semantical errors.
no, 53,3% impl.,
30,0%
expl., 16,7%
yes, 13,3% partly,
10,0%
no, 76,7%
yes, 40,0%
partly, 6,7%
no, 53,3%
Stroppi et al.
(2011), 6,7% BPMN
ext., 10,0%
own def. method,
6,7%
no, 76,7%
Definition Requirements Analysis Semantic Fit Check
Reuse of Artifacts Process Model
Valid Ext, 16,7%
None, 36,7%
Own Ext
Not., 10,0%
Own Ext, 36,7%
Fig. 3. Analysis of the extensions regarding their meta model definition and method-ological aspects
3.3 Applied Method
As already shown in section 3.2, the BPMN extension mechanism is rarely ap-plied. Nearly three out of four do not apply any method. These extensions aredeveloped in an ad hoc manner, what impedes the assessment of the replicabil-ity and comprehensibility. 16,7% of the extensions were designed based on theBPMN extension model (five in total), whereby only two applied the processmodel of [8]: [32] and [38]. [38] extends the process model concerning a semanticequivalence check to ensure the necessity of extension. Another two extensions
Classification of Domain-Specific BPMN Extensions 51
were designed based on individually outlined procedures [23], [39]. Regardingthe criterion of requirements analysis, approximately one of two articles providerequirements to the extension. One third was stated explicitly (e.g., by a set ofrequirements R1 to Rn, [20]). The rest of these articles describe requirementsimplicitly within the introduction or the description of the application context(e.g., [40]). Three of four articles designed the particular extension without anydeep consideration of the question, whether each requirement or extension de-mand needs necessarily an extension concept (see the middle piechart in figure3). 13,3% conducted a discussion for every concept [27], [33], [32], [38]. Further,nearly half of the extensions make use of existing domain artifacts. For instance,UML profiles [23], [39], domain modeling concepts [27], [32], [36] or requirements[30] are reused within the extension design.
Activities, 24,7%
Process, 15,3%
Common, 14,1%
Data, 14,1%
Artifacts, 9,4%
Gate- ways, 7,1%
Events, 5,9%
0,0%
2,0%
4,0%
6,0%
8,0%
10,0%
12,0%
Dat
a O
bjec
t Ta
sk
Act
ivity
P
roce
ss
Exp
ress
ion
Gat
eway
G
roup
R
esou
rce
Lane
S
eque
nce
Flow
A
rtifa
ct
Eve
nt
Par
ticip
ant
Per
form
er
Text
Ann
otat
ion
Tim
er E
vent
B
ound
ary
Eve
nt
Cho
reog
raph
y C
olla
bora
tion
Com
plex
Gat
eway
C
onne
ctin
g O
bjec
t D
ata
Sta
te
Dat
a S
tore
B
ase
Ele
men
t H
uman
Per
form
er
Man
ual T
ask
Par
alle
l Gat
eway
P
ool (
=Lan
e)
Pro
perty
R
esou
rce
Par
amet
er
Rol
e (R
esou
rce
Rol
e)
Ser
vice
Tas
k Ta
sk T
ype
Use
r Tas
k
BPMN Package BPMN Element
Others, 9,4%
Fig. 4. Distribution of extended BPMN packages and elements
3.4 Domain-specific Extension
Within the extension analysis, only publications with at least one identifiablenew element were considered. The number of new elements had a range be-tween one and 35 elements; on an average of nearly eight elements (e.g., [41],[24]). Examples for particular new elements as well as the derived size classescan be found in figure 6. Although the definition of new diagram types is notconsidered within BPMN, some extensions also provide the definition of newdiagrams like a Resource Structure Diagram [34] or a Secure Business ProcessDiagram [30], [31]. Next to the definition of new elements, BPMN elements arealso extended or customized. As figure 4 shows, primarily Data Objects, Tasks,Activities and Processes are extended. This fact is also emphasized by the pre-sentation of the extended BPMN packages: Elements from the Activity package,the Process package, the Common package (e.g., Resource, Sequence Flow orExpression) and the Data package are extended mainly. It could be concluded,that these elements are predestinated for domain-specific extensions. Especially,
52 R. Braun and W. Esswein
Data Objects and Tasks are often specified within extensions (see figure 6, col-umn “specialization”). The extension of standard BPMN elements is mainlyrealized by new relations (associations in the meta model) or specifications (in-heritances). Generally, the new relations are passive. It means that they are notmandatory from the perspective of the extended element but rather optional;extending the dynamic range of the referencing element. New relations betweenstandard elements are on rare occasions [24]. Also, the extension by owned at-tributes is implemented rarely (e.g., [32], [24], [40]), whereas the specificationof (generic) BPMN elements by new domain specific sub classes seems to be acommon means (e.g., [42], [43], [44], [29], [45]).
Meta model based
(abstract syntax), 62,5% Semantic,
7,1%
Graphical (concrete
syntax), 30,4%
0,0%
5,0%
10,0%
15,0%
20,0%
25,0%
30,0%
AS-A AS-Sp AS-En AS-A-B Sem-Co Sem-Ch CS-Dg CS-Cu CS-Co CS-Ah CS-Te
Extension Style Classes Extension Styles
Abstract Syntax (AS) Semantics (Sem) Concrete Syntax (CS)
Fig. 5. Distribution of identified extensions styles (56 in total over all extensions)
Consequently, the distribution of the applied extension styles reveals that theAS-Sp style (specialization) is one of the most applied extension techniques (seefigure 5). In total, 56 style applications could be identified within the extensiondefinitions. 62,5% of them affect the abstract syntax (meta model), 30,4% arerelated to the graphical notation (concrete syntax) and 7,1% realize some exten-sion by concrete or change some element’s semantics. Unsurprisingly, the (moreor less unspecific) AS-A style is applied the most. The enumeration technique(AS-En) for the domain-specific definition of ranges is used in more than 10% ofall style applications. Within the area of graphical style, there is no dominatingtechnique. Interestingly, CS-Co is not applied often, although BPMN explicitlyemphasize this possibility for artifacts elements [7].
4 Implications
Several implications can be derived based on the analysis of existing BPMN ex-tensions. We argue, that the following aspects should be considered in prospec-tive research works on BPMN extensions.
Strict Use of the BPMN Extension Mechanism: As shown, only veryfew extensions are designed by applying the BPMN extension mechanism. How-ever, such an implementation is indispensable for reasons of standard conformity,
Classification of Domain-Specific BPMN Extensions 53
N
ew e
lem
ents
Exte
nded
or c
usto
miz
ed e
lem
ents
Authors
Elements (sample)
Count
Size Class
New Diagram
Relations
Properties
Speciali-zation
Enhance-ment
Graphical Custom.
Extended Elements
Extension Style
Altu
hhov
et a
l. V
ulne
rabi
lity
Poi
nts,
Loc
k, A
ssoc
iatio
n Fl
ow
3 Ti
ny
- D
ata
Sto
re, D
ata
Obj
ect,
Task
-
- -
- 3
CS
-Cu,
CS
-Co,
AS
-A
Aw
ad e
t al.
O
rgan
izat
iona
l Rol
e, P
rofil
e, C
ase,
...
6 Li
ght
- Ta
sk, P
roce
ss, L
ane
Rol
e -
- 4
AS
-A, C
S-C
o B
occi
arel
li &
D'A
mbr
ogio
G
a W
orkl
oad
Eve
nt, A
rriv
al P
atte
rn, P
a Q
ualif
icat
ion,
...
11
Larg
e -
Col
labo
ratio
n, p
roce
ss (d
irect
ion
of
asso
ciat
ion
is n
ot d
efin
ed)
- -
- -
2 A
S-A
, CS
-Te
Bra
mbi
lla e
t al.
Soc
ial M
onito
ring,
Soc
ial B
ehav
iour
, UR
I, ...
14
La
rge
- -
- Ta
sk
- -
1 A
S-S
p B
raun
& E
ssw
ein
Hum
an R
esou
rce,
Mat
eria
l, A
lloca
tion
Sta
tes,
...
17
Larg
e -
Tim
er E
vent
, Par
ticip
ant
Dat
a O
bjec
t P
rope
rty, R
esou
rce
- -
5 A
S-A
, AS
-Sp,
AS
-En,
S
em-C
o B
raun
et a
l.
Med
ical
Doc
umen
t, C
PG
Ref
eren
ce, D
iagn
osis
Ta
sk, .
.. 14
La
rge
- A
ctiv
ity, G
roup
, Gat
eway
, Pro
cess
, Dat
a O
bjec
t -
Dat
a O
bjec
t, Ta
sk, P
aral
lel
Gat
eway
, Com
plex
Gat
eway
, E
xpre
ssio
n
- -
9 A
S-A
, AS
-Sp,
AS
-En,
S
em-C
o
Bru
cker
et a
l. A
cces
s C
ontro
l, S
epar
atio
n of
Dut
y, B
indi
ng o
f D
uty,
...
(4)
Tiny
- -
- -
- 0
AS
-A
Cha
rfi e
t al.
P
oint
cuts
, Adv
ice
(2)
Tiny
- -
- -
- 0
CS
-Cu
Frie
dens
tab
et a
l.
Dur
atio
n, P
roce
ss S
ectio
n, Q
uant
itativ
e Li
mit,
...
35
Hea
vy
(yes
) D
ata
Obj
ect,
Dat
a S
tate
, Pro
cess
, E
xpre
ssio
n, S
eque
nce
Flow
A
ctiv
ity,
Pro
cess
- -
- 6
AS
-A, A
S-S
p, A
S-E
n,
CS
-Dg
Gag
ne &
Tru
del
Tim
e D
ate
Exp
ress
ions
, Tem
pora
l Con
stra
int
Dec
orat
ors
(2)
Tiny
-
(Tem
pora
l Dep
ende
ncie
s, T
empo
ral
Con
stra
int A
ttrib
utes
) -
- -
0 A
S-A
Gro
ßkop
f P
erfo
rmer
Rol
e, P
roce
ssor
Rol
e, A
ctor
3
Tiny
-
- A
ctiv
ity
- -
- 1
AS
-A
Kop
p et
al.
TOS
CA
Nod
e M
anag
emen
t Tas
k, T
OS
CA
Scr
ipt
Task
, TO
SC
A D
ata
Obj
ect,
.. (4
) Ti
ny
- -
- Ta
sk, D
ata
Obj
ect
- -
2 A
S-S
p
Kor
herr
& L
ist
Wai
ting
Tim
e, W
orki
ng T
ime,
Pro
cess
Goa
l, ...
12
La
rge
- P
ool,
Pro
cess
, Gro
up, E
vent
, Tim
er
Eve
nt
- -
- -
5 A
S-A
-B
Lodh
i et a
l. M
etric
Val
ues,
Col
ors,
Rul
es/C
ondi
tion,
Pro
babi
lity,
D
imen
sion
al A
ttrib
utes
, Con
tent
/Stru
ctur
e 6
Ligh
t -
Act
ivity
, Gat
eway
s, C
onne
ctin
g O
bjec
ts,
Arte
fact
s, L
anes
-
- -
- 5
AS
-A, C
S-C
o, S
em-C
h
Lohm
ann
& N
yolt
Pla
ceho
lder
con
tain
er
(1)
Tiny
(y
es)
- -
Eve
nts,
Cho
reog
raph
ies
(onl
y gr
aphi
cally
!) -
- 2
CS
-Cu,
(CS
-Dg)
Mag
nani
& M
onte
si 2
007
Sim
ple
Cos
ts, C
ost I
nter
vals
, Ave
rage
Cos
ts
3 Ti
ny
- Ta
sk, G
atew
ay, A
ctiv
ity
- -
- -
3 C
S-A
h M
agna
ni &
Mon
tesi
200
9 S
tore
, Ent
ity, R
elat
ions
hip,
...
7 Li
ght
- -
- -
- D
ata
obje
cts
1 C
S-C
o
Mar
cink
owsk
i & K
ucia
pski
Ris
k Fa
ctor
, Ris
k Ty
pe, R
isk
Han
dler
9
Ligh
t -
Seq
uenc
e Fl
ow, R
esou
rce,
Par
ticip
ant
- A
rtefa
ct, T
ask
- -
5 A
S-S
p, A
S-E
n M
ülle
r-W
icko
p &
Sch
ultz
A
ccou
nt E
ntry
, Deb
it A
nd C
redi
t, A
ccou
nt, B
alan
ce
4 Ti
ny
- -
- D
ata
Obj
ect,
Gro
up, T
ext
Ann
otat
ion
- -
3 A
S-S
p
Nat
schl
äger
D
eont
ic T
ask,
O(T
ask)
, X(T
ask)
, P(T
ask)
, ...
(6)
Ligh
t -
- -
Task
-
- 1
- P
illat
et a
l. Ta
ilore
d B
ase
Ele
men
t 1
Tiny
-
- -
- -
- 0
AS
-Sp
Rod
rigue
z et
al.
Sec
urity
Req
uire
men
t, In
tegr
ity, P
rivac
y, ..
. 13
La
rge
yes
- -
Pro
cess
(Bus
ines
s P
roce
ss
Dia
gram
) -
- 1
AS
-A-B
, AS
-A, C
S-D
g
Sae
edi e
t al.
Rel
iabi
lity,
Cos
t, R
espo
nse
Tim
e, ..
. 7
Ligh
t -
Act
ivity
-
- -
- 1
AS
-A
Sal
eem
& H
assa
n A
cces
s C
ontro
l, A
uthe
ntic
atio
n, A
utho
rizat
ion,
...
8 Li
ght
yes
- -
- -
- 0
AS
-Sp,
CS
-Cu,
CS
-Dg
Sch
leic
her e
t al.
Com
plia
nce
Sco
pe, C
ompl
ianc
e R
ule,
Lan
guag
e In
dica
tor
(3)
Tiny
-
- -
- -
- 0
AS
-A-B
Stro
ppi e
t al.
Res
ourc
e P
rivila
ge, R
esou
rce
Bas
e, S
ubsu
mpt
ion,
...
17
La
rge
yes
Res
ourc
e, R
esou
rce
Par
amet
er,
Ele
men
t, B
ound
ary
Eve
nt, U
ser T
ask,
H
uman
Per
form
er, E
xpre
ssio
n
- -
- -
7 A
S-A
, AS
-En,
CS
-Dg
Sun
gur e
t al.
WS
N T
ask,
tWS
NO
pera
tion,
tWS
NP
erfo
rmer
, ...
7 Li
ght
- (R
esou
rce
Ass
ignm
ent)
Exp
ress
ion
- P
erfo
rmer
, Ser
vice
Tas
k -
- 3
AS
-Sp,
AS
-A, A
S-E
n S
upul
niec
e et
al.
Dat
a O
bjec
t, In
form
atio
n O
bjec
t, K
now
ledg
e
(3)
Tiny
-
Act
ivity
, Tas
k Ty
pe, R
ole
(Per
form
er),
Dat
a O
bjec
t -
- -
- 4
CS
-Ah,
(Sem
-Ch)
Wol
ter &
Sch
aad
Aut
horiz
atio
n C
onst
rain
t 1
Tiny
-
Man
ual T
ask,
Gro
up (n
ew re
latio
n, L
ane
- Te
xt A
nnot
atio
n -
- 4
AS
-A, A
S-S
p Zo
r et a
l. P
arts
Con
tain
er, M
achi
nes
and
Tool
s Fl
ow
Con
nect
or, .
.. 10
Li
ght
- -
- G
atew
ay, T
ask,
Res
ourc
e,
Seq
uenc
e Fl
ow
Act
ivity
-
5 C
S-A
h
Fig. 6. Analysis of the extensions regarding new and extended elements
54 R. Braun and W. Esswein
comprehensibility, model exchange and tool support. For instance, model engi-neers fail in reusing the most BPMN extensions since they do not provide avalid BPMN extension model. Thus, it is necessary to transform the provideddedicated meta model into a BPMN conform model in order to integrate itwithin a BPMN tool. Also, the communication within the research communityis hampered by this shortcoming. In the context of method engineering, it isalso necessary to define the concrete syntax of each extension element explicitlyto avoid misunderstandings. The semantics of a new element or its relations toBPMN elements should be described in detail in order to support its application.
Integrated Methodological Support Is Necessary: As stated at thebeginning of this article, BPMN lacks in term of providing an extension pro-cess model. Thus, most of the considered BPMN extensions are not designedrigorously. There seems to be a gap between the domain-specific definition ofextension requirements, their conceptualization and the implementation as validmeta model. The last aspect is successfully solved by [8] and few extensionsmake use of its proclaimed transformation procedure. However, the early phasesof extension planning and design are still not guided. Therefore, we see the needfor an integrated process model for BPMN extension development that focusesthe domain analysis and conceptualization phase. For example, there should bea systematic support for the decision whether any domain concept can be rep-resented within the “semantic scope” of a standard element or not. We suppose,that more than a few BPMN extensions do not exploit the entire expressivenessof BPMN. Besides, research on the integration of domain-specific artifacts withinBPMN extensions and DSMLs in general should be intensified, since such arti-facts (e.g., ontologies, taxonomies) provide well-defined domain knowledge thatcould complement domain expert knowledge.
BPMN Language and Extension Design: It became obvious, that spe-cific aspects are often demanded. Especially, a better resource and data objectmodeling needs to be supported by BPMN, albeit BPMN will not be understoodas any kind of a “data-flow language” ([7], p. 22). Referring to enterprise archi-tecture frameworks, an extension of the BPMN regarding several views (e.g.,resource perspective) is promising. Hereof, further research should consider thequestion, how to extend BPMN with new diagrams or views. Currently, such anextension is not designated. Also, based on our analysis of the so-called extensionstyles, a deeper analysis of extension patterns is necessary in order to providespecific patterns or guidelines for given extension purposes.
5 Conclusion
To the best of our knowledge, this is the first approach addressing the com-parison and classification of BPMN extensions in order to present the currentstate of the art. Therefore, an extensive literature review of BPMN extensionswas conducted that results in a set of 30 publications that were subjected anin-depth analysis. For the comparison and classification, a four-part extensionanalysis framework was designed containing criteria on the extension itself and
Classification of Domain-Specific BPMN Extensions 55
the applied procedures. Based on the application of this framework, several im-plications were derived.
First, authors of BPMN extension should strictly use the BPMN extensionmechanism in order to provide a valid extension and enable model exchange-ability (currently less than 20% provide valid BPMN extensions). Second, weidentified a need for an integrated methodological support of extension devel-opment, especially in terms of domain analysis and the comparison of domainelements with BPMN standard elements. Third, we have identified a recogniz-able need for the support of resource and data oriented modeling aspects withinBPMN. In this context, especially the question of extending BPMN (or a mod-eling language in general) with new diagrams and views should be considered.Regarding to the identified extension styles, it might be promising to developextension patterns (or at least guidelines) for specific extension needs in orderto support the design process.
References
1. Chinosi, M., Trombetta, A.: Bpmn: An introduction to the standard. ComputerStandards and Interfaces 34(1), 124–134 (2012)
2. ISO: Iso/iec 19510:2013 - International organization for standardization (iso)(2013)
3. Frank, U.: Outline of a method for designing domain-specific modelling languages.ICB Research Report 42, Universitat Duisburg-Essen, Essen (2010)
4. Mohagheghi, P., Haugen, Ø.: Evaluating domain-specific modelling solutions. In:Trujillo, J., Dobbie, G., Kangassalo, H., Hartmann, S., Kirchberg, M., Rossi, M.,Reinhartz-Berger, I., Zimanyi, E., Frasincar, F. (eds.) ER 2010. LNCS, vol. 6413,pp. 212–221. Springer, Heidelberg (2010)
5. Lammel, R., Ostermann, K.: Software extension and integration with type classes.In: Proceedings of the 5th International Conference on Generative Programmingand Component Engineering, pp. 161–170. ACM (2006)
6. Kopp, O., Gorlach, K., Karastoyanova, D., Leymann, F., Reiter, M., Schumm, D.,Sonntag, M., Strauch, S., Unger, T., Wieland, M., et al.: A classification of bpelextensions. Journal of Systems Integration 2(4), 3–28 (2011)
7. OMG: Business Process Model and Notation (BPMN) - Version 2.0. Object Man-agement Group (OMG) (2011)
8. Stroppi, L.J.R., Chiotti, O., Villarreal, P.D.: Extending BPMN 2.0: Method andtool support. In: Dijkman, R., Hofstetter, J., Koehler, J. (eds.) BPMN 2011.LNBIP, vol. 95, pp. 59–73. Springer, Heidelberg (2011)
9. Braun, R., Schlieter, H.: Requirements-based development of bpmn extensions –the case of clinical pathways. In: 1st International Workshop on the Interrelationsbetween Requirements Engineering and Business Process Management (2014)
10. Hevner, A., Chatterjee, S.: Design science research in information systems. In:Design Research in Information Systems. Integrated Series in Information Systems,vol. 22, pp. 9–22. Springer, US (2010)
11. Cooper, H.M.: Organizing knowledge syntheses: A taxonomy of literature reviews.Knowledge in Society 1(1), 104–126 (1988)
12. Vom Brocke, J., Simons, A., Niehaves, B., Niehaves, B., Reimer, K., Plattfaut, R.,Cleven, A.: Reconstructing the giant: on the importance of rigour in documentingthe literature search process. In: ECIS 2009 Proceedings, Paper 161 (2009)
56 R. Braun and W. Esswein
13. WKWI: Wi-orientierungslisten. Wirtschaftsinformatik 50(2), 155–163 (2008)14. Decker, G., Puhlmann, F.: Extending BPMN for modeling complex choreographies.
In: Meersman, R., Tari, Z. (eds.) OTM 2007, Part I. LNCS, vol. 4803, pp. 24–40.Springer, Heidelberg (2007)
15. Gao, F., Zaremba, M., Bhiri, S., Derguerch, W.: Extending bpmn 2.0 with sensorand smart device business functions. In: 2011 20th IEEE International Workshopson Enabling Technologies: Infrastructure for Collaborative Enterprises (WETICE),pp. 297–302. IEEE (2011)
16. Zor, S., Gorlach, K., Leymann, F.: Using bpmn for modeling manufacturing pro-cesses. In: Proceedings of 43rd CIRP International Conference on ManufacturingSystems, pp. 515–522 (2010)
17. Pardillo, J.: A systematic review on the definition of UML profiles. In: Petriu,D.C., Rouquette, N., Haugen, Ø. (eds.) MODELS 2010, Part I. LNCS, vol. 6394,pp. 407–422. Springer, Heidelberg (2010)
18. Natschlager, C.: Deontic BPMN. In: Hameurlain, A., Liddle, S.W., Schewe,K.-D., Zhou, X. (eds.) DEXA 2011, Part II. LNCS, vol. 6861, pp. 264–278. Springer,Heidelberg (2011)
19. Schleicher, D., Leymann, F., Schumm, D., Weidmann, M.: Compliance scopes:Extending the bpmn 2.0 meta model to specify compliance requirements. In: 2010IEEE International Conference on Service-Oriented Computing and Applications(SOCA), pp. 1–8. IEEE (2010)
20. Sungur, C.T., Spiess, P., Oertel, N., Kopp, O.: Extending bpmn for wireless sen-sor networks. In: 2013 IEEE 15th Conference on Business Informatics (CBI), pp.109–116. IEEE (2013)
21. Korherr, B., List, B.: Extending the epc and the bpmn with business process goalsand performance measures. In: ICEIS (3), pp. 287–294 (2007)
22. Awad, A., Grosskopf, A., Meyer, A., Weske, M.: Enabling resource assignmentconstraints in bpmn. Hasso Plattner Institute, Potsdam (2009)
23. Bocciarelli, P., D’Ambrogio, A.: A bpmn extension for modeling non functionalproperties of business processes. In: Proceedings of the 2011 Symposium on The-ory of Modeling & Simulation: DEVS Integrative M&S Symposium, Society forComputer Simulation International, pp. 160–168 (2011)
24. Friedenstab, J., Janiesch, C., Matzner, M., Muller, O.: Extending bpmn for busi-ness activity monitoring. In: 2012 45th Hawaii International Conference on SystemScience (HICSS), pp. 4158–4167. IEEE (2012)
25. Lodhi, A., Kuppen, V., Saake, G.: An extension of bpmn meta-model for evaluationof business processes. Scientific Journal of Riga Technical University. ComputerSciences 43(1), 27–34 (2011)
26. Magnani, M., Cucci, F.: BPMN: How much does it cost? An incremental approach.In: Alonso, G., Dadam, P., Rosemann, M. (eds.) BPM 2007. LNCS, vol. 4714, pp.80–87. Springer, Heidelberg (2007)
27. Altuhhov, O., Matulevicius, R., Ahmed, N.: An extension of business process modeland notation for security risk management. International Journal of InformationSystem Modeling and Design (IJISMD) 4(4), 93–113 (2013)
28. Brucker, A.D., Hang, I., Luckemeyer, G., Ruparel, R.: Securebpmn: Modeling andenforcing access control requirements in business processes. In: Proceedings of the17th ACM Symposium on Access Control Models and Technologies, pp. 123–126.ACM (2012)
29. Marcinkowski, B., Kuciapski, M.: A business process modeling notation extensionfor risk handling. In: Cortesi, A., Chaki, N., Saeed, K., Wierzchon, S. (eds.) CISIM2012. LNCS, vol. 7564, pp. 374–381. Springer, Heidelberg (2012)
Classification of Domain-Specific BPMN Extensions 57
30. Rodrıguez, A., Fernandez-Medina, E., Piattini, M.: A bpmn extension for the mod-eling of security requirements in business processes. IEICE Transactions on Infor-mation and Systems 90(4), 745–752 (2007)
31. Saleem, M., Jaafar, J., Hassan, M.: A domain-specific language for modelling secu-rity objectives in a business process models of soa applications. AISS 4(1), 353–362(2012)
32. Braun, R., Esswein, W.: Extending bpmn for modeling resource aspects inthe domain of machine tools. WIT Transactions on Engineering Sciences (87),450–458 (2014)
33. Großkopf, A.: An extended resource information layer for bpmn. Hasso-Plattner-Institute for IT Systems Engineering, University of Potsdam (2007)
34. Stroppi, L.J.R., Chiotti, O., Villarreal, P.D.: Extended resource perspective sup-port for bpmn and bpel. CIbSE, 56–69 (2012)
35. Magnani, M., Montesi, D.: BPDMN: A conservative extension of bpmn with en-hanced data representation capabilities. arXiv preprint arXiv:0907.1978 (2009)
36. Supulniece, I., Businska, L., Kirikova, M.: Towards extending bpmn with the knowl-edge dimension. Enterprise, Business-Process and Information Systems Modeling,69–81 (2010)
37. Zor, S., Leymann, F., Schumm, D.: A proposal of bpmn extensions for the manu-facturing domain. In: Proceedings of 44th CIRP International Conference on Man-ufacturing Systems (2011)
38. Braun, R., Schlieter, H., Burwitz, M., Esswein, W.: Bpmn4cp: Design and imple-mentation of a bpmn extension for clinical pathways. Research Report TU Dresden
39. Pillat, R.M., Oliveira, T.C., Fonseca, F.L.: Introducing software process tailoring tobpmn: Bpmnt. In: 2012 International Conference on Software and System Process(ICSSP), pp. 58–62. IEEE (2012)
40. Gagne, D., Trudel, A.: Time-bpmn. In: IEEE Conference on Commerce and En-terprise Computing, CEC 2009, pp. 361–367. IEEE (2009)
41. Charfi, A., Muller, H., Mezini, M.: Aspect-oriented business process modeling withao4bpmn. Modelling Foundations and Applications, 48–61 (2010)
42. Brambilla, M., Fraternali, P., Vaca Ruiz, C.K.: Combining social web and bpm forimproving enterprise performances: the bpm4people approach to social bpm. In:Proceedings of the 21st International Conference Companion on World Wide Web,pp. 223–226. ACM (2012)
43. Kopp, O., Binz, T., Breitenbucher, U., Leymann, F.: BPMN4TOSCA: A domain-specific language to model management plans for composite applications. In:Mendling, J., Weidlich, M. (eds.) BPMN 2012. LNBIP, vol. 125, pp. 38–52.Springer, Heidelberg (2012)
44. Lohmann, N., Nyolt, M.: Artifact-centric modeling using BPMN. In: Pallis, G.,Jmaiel, M., Charfi, A., Graupner, S., Karabulut, Y., Guinea, S., Rosenberg, F.,Sheng, Q.Z., Pautasso, C., Ben Mokhtar, S. (eds.) ICSOC 2011 Workshops. LNCS,vol. 7221, pp. 54–65. Springer, Heidelberg (2012)
45. Wolter, C., Schaad, A.: Modeling of task-based authorization constraints in BPMN.In: Alonso, G., Dadam, P., Rosemann, M. (eds.) BPM 2007. LNCS, vol. 4714, pp.64–79. Springer, Heidelberg (2007)